Professor Imanishi (WO 98/39352) and Professor Wengel (WO 99/14226) independently invented Locked Nucleic Acid (LNA) in 1997 and the first LNA monomer was based on the 2′-O—CH2-4′ bicyclic structure (oxy-LNA). This LNA analogue has since then showed promising results as antisense drug candidates. Other LNA analogues has also been synthesized exhibiting similar high affinity/specificity for example 2′-NH—CH2-4′,2′-N(CH3)—CH2-4′ (amino-LNA) (Singh, S. K.; Kumar, R.; Wengel, J. J. Org. Chem. 1998, 63, 10035-10039; Singh, S. K.; Kumar, R.; Wengel, J. J. Org. Chem. 1998, 63, 6078-6079), and 2′-S—CH2-4′ (thio-LNA) (Singh, S. K.; Kumar, R.; Wengel, J. J. Org. Chem. 1998, 63, 6078-6079, Kumar, R.; Singh, S. K et al. Biorg. Med. Chem. Lett. 1998, 8, 2219-2222). Large quantities of amino-LNA are crucial for its use in antisense. Scaling-up the previously described method of synthesis of amino-LNA has appeared to be difficult and encountered several major problems.
The first difficult reaction in the scale up work proved to be the regioselective benzylation of 3-O-benzyl-1,2-O-isopropylidene-4-C-hydroxymethyl-α-D-erythro-pentofuranose (Koshkin, A.; Singh, S. K.; Nielsen, P.; Rajwanshi, V. K.; Kumar, R.; Meldgaard, M.; Olsen, C. E.; Wengel, J. Tetrahedron 1998, 54, 3607-3630) (see FIG. 1, compound 1).
Working in the 100 g range the reaction yielded a product-mixture of compound 2, the 1′-benzylated and the di-benzylated material even under optimised conditions. The maximum yield of the desired compound 2 was 59% dropping to an average of 45-50% compared to 71% on smaller scale. Furthermore, compound 2 could only be isolated through tedious chromatography of closely eluting products.
The second key step in the original strategy causing problems during scale-up synthesis was the double nucleophilic substitution of the di-O-tosyl nucleoside 5 using benzylamine giving nucleoside 6 (Singh, S. K.; Kumar, R.; Wengel, J. J. Org. Chem. 1998, 63, 10035-10039). The reaction on larger scale (22 g) apparently afforded a second product identified as the oxy-LNA derivative. The desired N-benzylated-amino-LNA product 6 was obtained in only 15% together with 13% of the oxy-LNA by-product. For comparison, the reaction gives 52% of nucleoside 6 on a 8 g scale with no side reaction reported (Singh, S. K.; Kumar, R.; Wengel, J. J. Org. Chem. 1998, 63, 10035-10039).
Yet another problem encountered appeared to be the debenzylation of nucleoside 6 using ammonium formate and 10% Pd/C in methanol. It appeared to be only partial debenzylation as verified by mass spectroscopy, and the product 7 proved to be difficult to isolate from the reaction mixture.
The first synthesis of an oxy-LNA nucleoside was performed by a linear approach using uridine as starting material (Obika, S.; Nanbu, D.; Hari, Y.; Morio, J. A. K.; In, Y.; Ishida, T.; Imanishi, T. Tet. Lett. 1997, 38, 8735-8738) but by virtue of being a convergent synthesis the route developed by Wengel and coworker (Koshkin, A.; Singh, S. K.; Nielsen, P.; Rajwanshi, V. K.; Kumar, R.; Meldgaard, M.; Olsen, C. E.; Wengel, J. Tetrahedron 1998, 54, 3607-3630; Koshkin, A. A. et al., J. Org. Chem. 2001, 66, 8504-8512) became the method of choice for the synthesis of LNA nucleosides.
Amino- and thio-LNA was originally synthesised quite differently, but according to the present invention there are common intermediates that can be used for amino-LNA, thio-LNA, seleno-LNA, α-L-LNA as well as methylene-LNA at late stages in the overall synthesis.